Hi ! i have a weird doubt
I have seen some audio equipment with the regulators placed quite distant physically from the devices they supply
Could provide any benefit to add some local capacitance close to the output devices ?
if so how much is ok and how much instead is too much ?
Considering a dual power supply powering an opamp could two 470uF (or more?) caps on the + and - pins of the opamp help ?
I have look in the web and i have seen even 1000uF used in the case below
thank you very much and have a nice day
gino
I have seen some audio equipment with the regulators placed quite distant physically from the devices they supply
Could provide any benefit to add some local capacitance close to the output devices ?
if so how much is ok and how much instead is too much ?
Considering a dual power supply powering an opamp could two 470uF (or more?) caps on the + and - pins of the opamp help ?
I have look in the web and i have seen even 1000uF used in the case below
thank you very much and have a nice day
gino
Last edited:
The whole point of an active regulator is to respond very rapidly to changes in current demand (and so changes in regulated output voltage). If you add lots of capacitance you remove the ability of the regulator to do that as the cap becomes dominant and it slows the rate the output voltage can change and the also the rate the regulator can pump current into the load.
Smallish caps of up to around 47uF at most are all that is needed and many data sheets will specify even less. Decoupling the opamp at the pins is often useful for stability and these caps are usually very small such a 0.1uF as they are perfect to maintain a low impedance at higher frequency. I personally like to decouple from supply pin to supply pin rather than pin to ground at the opamp.
Smallish caps of up to around 47uF at most are all that is needed and many data sheets will specify even less. Decoupling the opamp at the pins is often useful for stability and these caps are usually very small such a 0.1uF as they are perfect to maintain a low impedance at higher frequency. I personally like to decouple from supply pin to supply pin rather than pin to ground at the opamp.
Hi Mr Mooly ! thank you very much for your always kind and very valuable advice
It could be that a good quality capacitor provides a much lower impedance than the one of the regulator ?
In photo flash where the speed of discharge is everything there is a high voltage cap providing the instant power required by the flash
I see music peaks as flashes And i have the feeling that the discharge of a cap is much faster than the regulator
Expecially when the caps are placed directly on the output devices pins There will be no lag in the delivery of current
The regulator will have only to refill the cap
Anyway i would look at the esr of the cap Taken a same series caps with higher uF will have a lower ESR
Low ESR should mean fast discharge and fast refilling
I am talking of layouts where the regulators are quite distant from the opamps
I understand it could be a debatable design choice
Very helpful indeed I do not know it it is relevant but i wonder how the regulator impedance value changes with frequency
It could be that a good quality capacitor provides a much lower impedance than the one of the regulator ?
In photo flash where the speed of discharge is everything there is a high voltage cap providing the instant power required by the flash
I see music peaks as flashes And i have the feeling that the discharge of a cap is much faster than the regulator
Expecially when the caps are placed directly on the output devices pins There will be no lag in the delivery of current
The regulator will have only to refill the cap
i see I do not even have a clue about the max current draw of an opamp actually And moreover if the circuit works in class A the current draw will be constant In this sense class A circuit eliminates the problem of impedance because the current flow is constant This is very good
Anyway i would look at the esr of the cap Taken a same series caps with higher uF will have a lower ESR
Low ESR should mean fast discharge and fast refilling
I am talking of layouts where the regulators are quite distant from the opamps
I understand it could be a debatable design choice
The regulator output impedance rises with frequency which is why a suitable small cap is needed to maintain a low impedance as frequency rises.
Opamps in the scheme of things draw little current and the peak load currents are also low and so there is no need for large reserves of current.
In audio it is all about maintaining a constant supply voltage with the minimum of noise present.
Think of the cap in the regulator as being like a car pulling a trailer. The speed is like the voltage output of the regulator. The engine and brakes represent the ability to speed up and slow down the car. If the trailer is very heavy (a big cap) the car cannot respond quickly to increasing and decreasing the speed. If the trailer is light (a small cap) it can respond much faster to changes.
Opamps in the scheme of things draw little current and the peak load currents are also low and so there is no need for large reserves of current.
This is different thing altogether. The energy in the cap is dumped into the flash tube when the tube is triggered. Use a bigger cap and the energy delivered to the tube could be destructive.
In audio it is all about maintaining a constant supply voltage with the minimum of noise present.
Think of the cap in the regulator as being like a car pulling a trailer. The speed is like the voltage output of the regulator. The engine and brakes represent the ability to speed up and slow down the car. If the trailer is very heavy (a big cap) the car cannot respond quickly to increasing and decreasing the speed. If the trailer is light (a small cap) it can respond much faster to changes.
Good morning Mr Mooly, thank you very much again for your kind and precious support I understand things much better now
I got an old AMC DAC8 from a friend emptying his garage
I am attaching the schematic and a picture of the interiors
i removed the mains transformer and placed a dc panel socket on the back and i am using an external laptop DC power supply
On the blue board on the back there is a 7815 that i have kept
I am thinking to bypass it and supply the dac directly with an external 15VDC regulated linear power supply i already have at hand
This is a picture of the regulator with new caps around
better to keep the regulator of remove it ?
I got an old AMC DAC8 from a friend emptying his garage
I am attaching the schematic and a picture of the interiors
i removed the mains transformer and placed a dc panel socket on the back and i am using an external laptop DC power supply
On the blue board on the back there is a 7815 that i have kept
I am thinking to bypass it and supply the dac directly with an external 15VDC regulated linear power supply i already have at hand
This is a picture of the regulator with new caps around
better to keep the regulator of remove it ?
I would keep the 7815 regulator and if you are feeding that from an external DC supply of any kind then it needs to be at least 19 volts DC to give the 7815 enough headroom to work with.
Using an external 15 volt supply and removing the 7815 will work but I would not do that. Better to keep the regulator in the box than have long wires to the DAC from an external PSU.
Using an external 15 volt supply and removing the 7815 will work but I would not do that. Better to keep the regulator in the box than have long wires to the DAC from an external PSU.
Hi thank you very much again
Honestly i was thinking to place some more caps on the board to create a kind of energy reservoir
Maybe i have been confused too much by reading people bashing the single chip regulators and pushing for the use of discrete regulators
But I will do as you say
still i do not understand why in some cases regulators are so far from the circuits they supply
Expecially in dacs and preamps of course
Power amps are a different application
Thank you sincerely again
Honestly i was thinking to place some more caps on the board to create a kind of energy reservoir
Maybe i have been confused too much by reading people bashing the single chip regulators and pushing for the use of discrete regulators
But I will do as you say
still i do not understand why in some cases regulators are so far from the circuits they supply
Expecially in dacs and preamps of course
Power amps are a different application
Thank you sincerely again
Last edited:
There is no need because the current draw will be very constant in a DAC. If it was not constant (but it will be with this) then adding caps destroys the regulators ability to respond quickly.
That's a good question. Given the current draw is fairly constant there is no real problem with the supply varying due to changing load currents. The board will have lots of local (small) decoupling caps to maintain a low impedance.
Hi ! thanks a lot for the very kind and valuable advice I read NJM2860 in the schematic
is there a good alternative that does not imply a circuit modification ?
is there a good alternative that does not imply a circuit modification ?
LM317 Output cap value.
Look at the data sheets:
1 micro farad ( or 10 micro farad ).
I see they put 47uF for a high current regulator using an LM317 plus a power BJT and driver BJT.
One should never use a componant not reading, understanding, following the datasheet. And application notes.
Most if not all the debugging I have done in my life has been on hardware put together with no care about the datasheets.
Most fixes were just: Correct implementation from the datasheet.
Look at the data sheets:
1 micro farad ( or 10 micro farad ).
I see they put 47uF for a high current regulator using an LM317 plus a power BJT and driver BJT.
One should never use a componant not reading, understanding, following the datasheet. And application notes.
Most if not all the debugging I have done in my life has been on hardware put together with no care about the datasheets.
Most fixes were just: Correct implementation from the datasheet.
Last edited:
Hi ! thank you for your kind and precious advice
actually i looked at the datasheet and i noticed the very low values The original output cap was very small indeed
I put a 470uF one because i did not have smaller values
I will replace it with a 10uF one I hope i did not ruin the regulator
thanks a lot again
actually i looked at the datasheet and i noticed the very low values The original output cap was very small indeed
I put a 470uF one because i did not have smaller values
I will replace it with a 10uF one I hope i did not ruin the regulator
thanks a lot again
In the case of the AMC DAC8 that you posted photos of, it looks like it has been done for modularity. This could make it easier to fabricate, or more repairable.
Hi ! thank you for your kind and helpful advice I will try a linear 20VDC power supply soon I am using smps intended for laptops
Still i read a thread about passive filtering of smps but i cannot find it anymore
Fwir a quite simple combination of caps of different values in parallel was having a nice suppression effect on noise
Still i read a thread about passive filtering of smps but i cannot find it anymore
Fwir a quite simple combination of caps of different values in parallel was having a nice suppression effect on noise
Yes it might be essential for stable operation of the load.
Depends on the load and the regulator - the regulator might have limits on maximum output capacitance for stability for instance, different loads may have different decoupling requirements.
An opamp doesn't need anything like 470µF, 10µF is probably OK with most regulators. Its not a great idea to mount linear regulators remote to the load, better for a +/-15V system to have remote power come in at +/-18V and use point-of-load regulation down to +/-15V. Basically on the same PCB.
Most modern opamps also absolutely require high-speed decoupling, something like 100nF ceramic say, very close to every chip, for correct operation - this is essentially independent of the need for regulation of the supply rails.